[ENH] Shapelet transforms vary in class attributes

[IRKnyazev]

Describe the feature or idea you want to propose

Overarching idea is to make the shapelet transforms implemented more similarly to one another, but for now they should be made to have the same class attributes for example shapelet transform has useful class attributes such as (shapelet information gain, shapelet length, start position the shapelet was extracted from, shapelet dimension, index of the instance the shapelet was extracted from in fit, class value of the shapelet. These are all very helpful for shapelet interpretation but aren't always present in the other transforms.

Describe your proposed solution

Add missing class attributes to SAST,RSAST (RDST has already been done I think). Though these attributes may not serve any purpose in the functionality of the algorithms, hence why they probably weren't added to the class, they are useful for comparing the shapelets generated by each transform,

Describe alternatives you've considered, if relevant

No response

Additional context

No response